Lettering

The Dean dislikes letters of recommendation (and has disliked them previously).  I agree: they've devolved into mostly useless exercises of hysterical hyperbole, in which every candidate is the Absolute Bestest Ever, Author of the Most Awesome Dissertation in the Known Galaxy, Capable of Teaching Three Hundred Freshmen in a Single Bound.  When it comes to letters of rec,  postmodernism reigns supreme; after all, given the aforementioned hysterical hyperbole (Absolute Bestest Ever, etc.), readers seek out silences.   No mention of a candidate's teaching in any letter? Letters from everyone except the dissertation director, who doesn't have the convenient excuse of being dead? No letters dated more recently than ten years ago? A passel of vaguely cheery, entirely nonspecific, unbelievably short letters for a "well-liked" longtime colleague?

The exceptions to this rule are the letters loaded with specifics.  The dissertation will remake the field, because...  The student stands out from the others I've taught in the past five years, because... This scholar's book is revolutionary, because...  Letters that transcend the boilerplate (X is a brilliant, dedicated student who has written a fine dissertation, teaches impeccably, and babysits my temperamental Persian cats without a sign of complaint) do make a positive impression.  Otherwise, they tend to melt into a pool of textual goo. 

(Incidentally, I'm fascinated by the second comment to the first of the Dean's posts.  That takes the Meaningless Death of Trees to new heights.)

UPDATE: Hmm.  I was clearly being hysterically hyperbolic when I wrote this last night, but it remains true that, in my experience, letters have never changed our search committees' minds--only confirmed an already positive or negative impression.  (If your letters were written ten years ago, for example, the cover letter generally and unintentionally indicates why.)